This invention relates generally to error processing in fault tolerant computing systems and specifically to a method of processing certain errors through software.
Computer resource overhead in software error processing slows down the operation of a computer system. A completely robust system which conducts a complete software check and analysis of all errors detected will be extremely inefficient. Additionally, important operations performed by the computer system can be delayed excessively or frustrated because of excessive time spent executing data processing operations to recover and locate a fault.
Despite the problems of such delay, many conventional error processing schemes require the computer system to suspend normal data processing operations and execute data processing operations for every error detected. Such an operation is wasteful.
On the other hand, some software analysis is often required to ensure that all errors are properly handled. Proper error handling generally requires logging of the error, locating the source of the error, if possible, and resuming operation, if possible.
It is advantageous to minimize the interruption to normal data processing operations by software error handling procedures.
It is also advantageous for the present invention to ensure proper use of software error handling procedures to maximize system efficiency.